Polaris Installation and deployment

graph TD;
subgraph Polaris
Z[Docker Build]
end
Z[Docker Build] --> A(Ingress);
Z[Docker Build] --> B(Ingress);
subgraph GitLab
B[Container Registry] --> H(Images);
end
subgraph Server
A[Docker Pull] --> D(Traefik);
D[Traefik] --> E(Deployment:Analytics-Engine);
D[Traefik] --> F(Deployment:Rights-Engine);
D[Traefik] --> I(Deployment:Dashboard-SDK);
E(Deployment:Analytics-Engine)---id1[(MongoDB)]
F(Deployment:Rights-Engine)---id1[(MongoDB)]
I(Deployment:Dashboard-SDK)---id1[(MongoDB)]
end

Getting started

Each folder contains one of the components of

Local Deployment

Login in the docker registry

If you are not already logged in, you need to authenticate to the Container Registry by using your GitLab username and password. If you have Two-Factor Authentication enabled, use a Personal Access Token instead of a password.

docker login registry.git.rwth-aachen.de

Create Docker network

Different containers communicate via a Docker network web, which must be created before Docker compose files are executed.

docker network create web

Create private/public keys and .env

Please change all passwords in the configuration file and leave the passphrase empty.

ssh-keygen -b 4096 -f id_rsa
cp .env.sample .env

Start containers

docker compose up -d

Please check whether all services started successfully.

docker compose ps

After that you should be able to visit http://localhost:8004/ and see the rights engine.

Migrate rights-engine database (only required once after first start)

docker compose exec -it rights-engine sh -c 'python3 manage.py sqlflush | sed s/TRUNCATE/DROP\ TABLE\ IF\ EXISTS/g | python3 manage.py dbshell && echo DROP\ TABLE\ IF\ EXISTS\ django_migrations\; | python3 manage.py dbshell && python3 manage.py migrate && python3 manage.py loaddata fixtures/initial_db.json'

Migrate analytics-engine database (only required once after first start)

docker compose exec -it scheduler sh -c 'scheduler create-db'

Adding analytics engine jobs

Analytics engines jobs are configured via yml files and read from the configuration directory, which is a volume.

Example configuration file configuration/h5p_engines.yml

h5p_statements_count_engine:
    crontab: "*/1 * * * *"
    repo: "https://git.rwth-aachen.de/polaris/public/engines/activities.git"
    analytics_token: "b6a4ec069ef9f688e781161d46c2a85c14a761a4eaf0074099656c7de44a65d9"

Example configuration file configuration/moodle_engines.yml

moodle_statements_count_engine:
    crontab: "*/1 * * * *"
    repo: "https://git.rwth-aachen.de/polaris/public/engines/activities.git"
    analytics_token: "0482a0f3259c966dfddb38de867573a95995ee5e10898bb71f9ae0e99efe54ef"

Update analytics engine scheduler

docker compose exec -it scheduler sh -c 'scheduler read-configs'

Create visualization token

curl -X POST http://localhost:8004/api/v1/provider/visualization-tokens/create --data '{"engines": ["count_h5p_statements", "count_moodle_statements"]}' -H "Content-Type: application/json"

Returns JWT Token for dashboard

{"token":[JWT_TOKEN]"}

Loading JSON statements

It is recommended to import a set of sample statements into the LRS so that the analytics engines can work on this data. Furhtermore, users (user1@polaris.com and user2@polaris.com) can test the data disclousure and data deletion process.

$ docker compose exec -it mongodb_container sh -c 'mongoimport --authenticationDatabase admin --username root --password CHANGE_ME --db lrs --collection statements --drop --file ./lrs_statements_dump.json'

Start dashboard

1. Clone
   • cd dashboard-example
2. Download latest @polaris/dashboard-sdk-X.X.X.tgz from registry and npm install @polaris/dashboard-sdk-1.0.2.tgz it (TODO: improve with npm login)
3. Update TOKEN in dashboard-example/src/js/app.js
4. Run npm run dev
5. Visit http://localhost:8005/

(Optional) Filling DB with random statements

1. Clone rights-engine
2. Create provider config

$ cd rights-engine/tools/xapi-statement-generator
$ cp provider_config.json.example provider_config.json

3. Open provider_config.json and insert Application Tokens (visible at http://localhost:8004 (Rights Engine UI), if you login as a provider)
4. Run python generator.py -t http://localhost:8003/xapi/statements -r

Update Docker Images

docker compose pull